JP3057496B2 - Electronic fuel injection control method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Summary] One of the factors that determine the fuel injection amount includes a coefficient FAF that performs feedback control of the air-fuel ratio, and learns the change over time of the coefficient FAF as another factor. In the case where a learning term to be corrected is included, when it is determined that the learning term is abnormal, the learning term is corrected to a value close to the initial value to prepare a condition for enabling feedback control.

[Industrial applications]

The present invention relates to an electronic fuel injection control device that introduces a learning term into control of an air-fuel ratio.

[Conventional technology]

In a fuel injection type internal combustion engine, the air-fuel ratio (A / A
F) depends on the fuel injection amount. In order to maintain this air-fuel ratio at the stoichiometric air-fuel ratio (λ = 1), the oxygen concentration in the exhaust gas is detected by an O ₂ sensor, and the mixture supplied to the internal combustion engine is rich (RiCH: excessive fuel) or lean (LEAN). : Fuel shortage) and feedback to correct fuel injection time (F / B)
Control is often performed.

FIG. 4 is a system configuration diagram of this type of engine control device, wherein 1 is an engine, 2 is a spark plug, 3 is an injector, 4 is a throttle, 5 is a starter, and 6 is an engine control computer. This computer 6 has an input interface, A / D converter, CPU, memory,
It comprises an output interface and the like, and outputs an ignition control signal for the spark plug 2 and an injection control signal for the injector 3. Inputs include the intake air temperature THA from the sensor 7, the intake pipe pressure PM from the sensor 9, the water temperature THW from the sensor 10,
Rotational speed NE from the sensor 11, in addition to the O ₂ concentration of the O ₂ sensor 12, there is a throttle opening degree.

 The fuel injection time TAu is set, for example, as follows.

TAu = (TP + TAuG) * FAF ... (1) or The basic injection time TP is mainly determined by the intake pipe pressure PM and the rotational speed NE. The output of the O ₂ sensor 12 is inverted depending on whether the O ₂ concentration in the exhaust gas is rich or lean with respect to the stoichiometric air-fuel ratio. At this time, the correction term FAF based on the air-fuel ratio A / F changes as shown in FIG. This change in FAF is the result of F / B control that increases fuel when lean and decreases fuel when rich.
However, in order to speed up the response at the time of inversion A, B, C...

The F / B control is stopped when the O ₂ sensor 12 is inactive or the like, and the FAF at this time is fixed to a predetermined value (for example, 1.0). On the other hand, the average value (F / B center) of the FAF during the F / B control is initialized so as to be the same value. However, it is sufficiently predicted that the F / B center deviates from a predetermined value due to a decrease in the pressure of the fuel pump or the injector over time, etc., so a learning term TAuG, KG is introduced, and the F / B center is set to a predetermined value. The learning term is shifted from the initial value 1.0 by an amount deviated from the value, and as a result, the F / B center is maintained at a predetermined value (1.0 in this case) so that the stoichiometric air-fuel ratio is maintained even during F / B non-control.

As a learning method, for example, every time the FAF is inverted, the average value (A + B) / 2, (B + C) / 2, (C + D) / 2,...

(I) When the average value is greater than 1.02, the learning value is increased by 0.1%.

(Ii) When 0.98 ≦ average value ≦ 1.02, the learning value is not updated.

(Iii) If 0.98> average value, decrease the learning value by 0.1%.

Therefore, in the case of equation (1), the basic injection time TP is TP + TAuG according to the actual system as shown in FIG.
₁ is modified into (rich side) or TP-TAuG ₂ (lean side), (2) as shown in FIG. (B) in the case of type TP * KG ₁
(Rich side) or TP * KG ₂ (lean side). At this time, the range in which F / B control is possible using the FAF is a hatched area.

[Problems to be Solved by the Invention] The learning function described above functions effectively even in the event of a system abnormality (such as an abnormality in the intake air temperature sensor or the water temperature sensor). However, if the learning term deviates greatly when the system abnormality is recovered, as shown in FIG.
It may not reach the TP at which F (λ = 1). For example, TP = 1200 (μS), When TAuG = 500, FAF = 0.8~1.2, (TP + TAuG 1) * FAFmin = (1200 + 500) * 0.8 = 1360> TP (TP + TAuG 2) * FAFmax = (1200 + 500) * 1.2 = 840 <TP, and λ = 1 cannot be reached.

In such a case, the output of the O ₂ sensor does not reverse because the F / B control is not performed normally as shown on the right side of FIG. As a result, the FAF averaging process and learning control are not performed, so that the average value of the FAF remains largely deviated from a predetermined value, causing the engine to become unstable due to the A / F deviation, exhaust gas failure, catalyst overheating. , Causing engine stalls and other obstacles.

The present invention is intended to solve the above-described problem by adding a function of correcting a learning term to an initial value.

[Means for solving the problem]

The present invention sets a fuel injection amount based on output values from various sensors, installs an O ₂ sensor in exhaust gas of an internal combustion engine to determine an air-fuel ratio of a supplied air-fuel mixture, and converts the air-fuel ratio to a theoretical value. Electronic fuel injection control that introduces a learning amount (TAuG, KG) that corrects the fuel injection amount so that the average value of the feedback control coefficient (FAF) becomes a predetermined value while performing feedback control of the fuel injection amount so as to maintain it. In the device, when it is detected that a difference between a learning amount stored before the ignition switch is turned off and a learning amount calculated from an initial value that is a predetermined value after the ignition switch is turned on is equal to or more than a predetermined value. And resetting means for directly using the learning amount calculated from the initial value after the ignition switch is turned on.

[Action]

When the difference between the learning term stored before the ignition switch is turned off and the learning term calculated from the initial value after the ignition switch is turned on is equal to or greater than a predetermined value, the learning term calculated from the initial value after the ignition switch is turned on. If reset means are used as is, learning control appropriate for the vehicle environment can be performed without complicated calculations, for example, when replacing sensors, and operation can be performed while the air-fuel ratio deviates from the theoretical value. It can be prevented from continuing.

〔Example〕

FIG. 1 is a flowchart of the first embodiment of the present invention. In this example, the learning amount is reset (RESET) when the system recovers normally from a system failure (for example, a sensor failure) using the system failure diagnosis function of the main routine. Step S1 is the calculation of the learning amount, which is performed as described above. Some processing is performed on the way, and a system failure diagnosis is performed in step S2. If an abnormality is detected here, it is stored in the memory (steps S3, S4). Then, the injection amount (TAu) is calculated and injected (steps S5 and S6). If it is normal in step S3, it is determined in step S7 whether or not the previous time is normal based on the contents of the memory. If the previous time is normal, the process directly proceeds to step S5.
In step S8, the learning amount is reset.

At this time, if a diagnostic code of “learning value abnormality” is stored at the same time, it is useful for vehicle maintenance. The learning value RE
SET returns to an initial value (for example, 1.0) or a method of reducing the abnormal learning value at that time to 1 / n (n is sequentially increased to 2, 3, 4,...).

FIG. 2 is a flowchart of the second embodiment of the present invention. In the present example, the learning amount is reset by detecting a characteristic shift such that the sensor system does not determine the failure. Complete warm-up for this, O ₂ (not an O ₂ sensor fail) sensor activity by in F / B region, the signal of the O ₂ sensor is not inverted over a predetermined time T, to detect, at this time resetting since recovery is impossible in the case where the learning value is on the same side as the output of the O ₂ sensor.

Step S11~S14 is part of the main routine, step S11 determines the F / B region, step S12 outputs the inversion detection of O ₂ sensors, step S13 is clear of the counter CNT.
This counter CNT is automatically counted by another routine.

Steps S20 to S29 are a system abnormality determination routine. Step S20 is the determination of the F / B area, and step S21 is the determination of the learning area. If not, the counter CNT is cleared in step S22. In the F / B area and the learning area, the value of the counter CNT is set to a predetermined time T in step S23.
Compare with Because if CNT> T O ₂ sensor is not reversed over time T, O ₂ sensor output at step S24 RiCH
Check if is LEAN. If it is RiCH, the learning amount is checked in step S25, and if it is RiCH side, the learning amount is calculated in step S26.
ESET. If the O ₂ sensor output is LEAN, step S27
To check the learning amount, and if it is also on the LEAN side, step S26
To reset the learning amount. In contrast, O ₂ sensor output and the learning of RICH, is better not to reset the learning amount when the relationship LEAN is reversed, the control on good learning amount is left intact. Steps S28 and S29 are the same as steps S5 and S6 in FIG.

FIG. 3 is a flowchart of the third embodiment of the present invention. In this example, the learning amount TAuGo stored before iG (ignition) OFF and the learning amount calculated from the initial value after iG ON
The abnormality is determined from the difference from TAuGi. Step S
Reference numeral 31 denotes an initial routine, which is a process for reading TAuGi, which is a readout of the learning amount stored before the iG is turned off, into TAuGo of the following equation. Steps S41 to S47 are a main routine. Step S41 is the calculation of TAuGi, and the calculation of the learning amount is started from an initial value that is a predetermined value, for example, 1.0. Step S42 is a process for judging the stability of F / B. In this process, 0.98 ≦ FAF (average) ≦ 1.02 in the same operation range,
It is determined whether FAF skipping has been performed n times or more (n is a number that reaches the guard assuming that learning has progressed only in the negative direction) or more in the same operation range. If it is determined that the F / B is stable, the magnitude of the difference between TAuGo and TAuGi is determined in step S43. In particular Is determined to be equal to or higher than a certain level. This level is, for example, a value of 30% or more.

If it is> level, TAuGi is used as it is in step S44. That is, TAuGo before iG OFF is reset.
On the other hand, if ≤ level, the difference is small, and therefore, in step S45, TAuGo is replaced with TAuGi. Steps S46 and S47 are the same as steps S28 and S29 in FIG.

〔The invention's effect〕

As described above, according to the present invention, the abnormal learning amount is automatically corrected to the initial value side, so that it is possible to prevent the operation from continuing with the air-fuel ratio deviating from the theoretical value, and to reduce the engine speed. Stability (idle, running surge, car knock, etc.) is improved. Further, overheating of the catalyst, exhaust gas failure, engine stall, knocking, and the like can be prevented. Furthermore, in this application, by reflecting the learning terms after iG is turned on,
For example, replace the sensor with a normal sensor before iG is turned on, and when returning to normal after iG is turned on, use what has been learned from iG on as it is, without performing complicated calculations and always matching the vehicle environment. Learning control.

[Brief description of the drawings]

1 to 3 are flowcharts showing different embodiments of the present invention, FIG. 4 is a configuration diagram of an engine control system, FIG. 5 is an explanatory diagram of feedback control, FIG. 6 is a characteristic diagram of injection time, FIG. FIG. 7 is an enlarged view of FIG. 6 (a).

Claims (1)

(57) [Claims] A fuel injection amount is set based on output values from various sensors, an O ₂ sensor is installed in exhaust gas of an internal combustion engine to determine an air-fuel ratio of a supplied air-fuel mixture, and a logical value of the air-fuel ratio is determined. The electronic fuel injection which introduces a learning amount (TAuG, KG) for correcting the fuel injection amount so that the average value of the feedback control coefficient (FAF) becomes a predetermined value while feedback-controlling the fuel injection amount so as to maintain the fuel injection amount. The control device detects that a difference between a learning amount stored before the ignition switch is turned off and a learning amount calculated from an initial value which is a predetermined value after the ignition switch is turned on is equal to or more than a predetermined value. Then, the electronic fuel injection control device further comprises reset means for directly using the learning amount calculated from the initial value after the ignition switch is turned on.